<headius[m]>
byteit101: if you are around today I would like to talk through remaining questions
<headius[m]>
vortex1111: perhaps you could start a branch to swap things around and we can assist as you go? LIke you say it should mostly be a search and replace if the APIs are the same
<headius[m]>
or the other option, like kares suggested, would be finding a shim library
<headius[m]>
boc_tothefuture: I think this is a terminology issue... that article is referring to java.util.Enumeration
<headius[m]>
if you want to use each on a Java `enum` then just call values first
<headius[m]>
kalenp: jswenson on the two issues we have open now...
<headius[m]>
json issue I will be raising to OpenJDK folks since it seems to be G1 issue or jit issue based on Freaky findings
<headius[m]>
tweaking JIT or GC flags may be a workaround for you in the interim
<headius[m]>
the memory/metaspace issue is still lacking information but we did get some feeback from Attila Szegedi of Nashorn... so perhaps on your side you could describe how Nashorn is being used versus JRuby and we can start to sort out if it is really a JRuby issue
<headius[m]>
Freaky: thank you for the updates... I'm going to post this in one of the openjdk mailing lists and then look at opening an issue
<boc_tothefuture[>
headius: Got it.. I ended up doing that, I misread it. thanks.
<jswenson[m]>
headius: I’m doing a little research to fully understand how we’re using nashorn. Will update the ticket when I have a better handle on it.
<headius[m]>
yeah thanks... I just don't see a lot of evidence that the LambdaForm explosion is our fault... that isn't to say we aren't eating up a lot of metaspace but without full indy available it wouldn't be due to LFs
<kalenp[m]>
wrt the json issues, we've rolled back to java8 for our most affected customers until we get more information, since we don't have a good enough repro to validate any active mitigations. I'll keep monitoring here for updates on the jit/gc issues
<headius[m]>
they are not great about responding so I have also pinged some friends at Oracle and will open an issue soon
sdlin[m] has joined #jruby
<Freaky>
\o/
<kalenp[m]>
@sdlin who just joined is another Looker, who's been working on our json parsing issue a bit more actively. he's going to pick up primary communication on that issue for us, so we have one fewer hop in the game of telephone
<headius[m]>
Alexsey is pretty good about jumping on JRuby issues
<headius[m]>
moving right along
<jswenson[m]>
@sdlin
<jswenson[m]>
sdlin: I wonder if the JDK9 fix may be related to what we’re seeing
<byteit101[m]>
headius: I'm around, though I may be busy at times and take a moment to get back to you
<headius[m]>
ok gonna hit gym for a bit and will get back into it after
<headius[m]>
it would help if you can review your own questions and filter out anything that doesn't need an answer anymore
<headius[m]>
I feel like we were retreading some ground
<byteit101[m]>
The remainder all look good except for "CJP is a mess because I don't know how to install the proper handlers, see that in General above", which I guess will now stay that way
<headius[m]>
look good as in they are still open questions?
<byteit101[m]>
correct
<headius[m]>
ok
<byteit101[m]>
(sorry :-) I know this is a lot of questions)
<headius[m]>
no problem I just want to make sure to focus my time right